Summary

This book covered the basic statistic concepts to use ML models. We can group our data to research the segments that are more important for our activity. Then, we can apply the regression model to these segments and see what the most influential variables are to build predictions using them. Finally, we forecast the values of these important segments and have an idea of what the behavior of our research variables can be in the future.

In this chapter, we applied the three general steps of the ML algorithm by designing a forecast model with known data, testing the model, also with known data, and finally, doing a prediction. As with any other ML function, we need to observe and ensure quality control of the data source to be sure that it is useful to do forecasts. In the case of time series, we know that the Durbin-Watson test checks for data autocorrelation to see whether the past has an influence on the present and whether it can predict the future. Then, we test the model...